Project Summary Vision is the sensory modality that humans depend on most to navigate through the world and to establish meaningful social relationships. Disorders of the visual system therefore often lead to severe interpersonal and economic hardships. Although much progress has been made dissecting the retinal mechanisms that generate action potentials in response to light, surprisingly little is known about how the brain actually interprets and contextualizes these signals. I propose to leverage cutting-edge optical technologies to clarify how neuronal ensembles (coactive groups of cells) encode sensory information in primary visual cortex (V1), the neocortical region which is required to perform cognitively demanding tasks including visual discrimination. Ensembles in V1 exhibit activity patterns with reproducible spatial and temporal structures which define the functional vocabulary of cortical microcircuits and that likely carry information pertaining to both the visual field and to the internal state of the animal. Yet how these computations emerge, whether they are comprised of independent circuits, and the extent to which they provide a template for visual perception remain unknown. To address these questions, I will perform chronic two-photon holographic calcium imaging of a consistent population of V1 neurons in awake mice learning a visually-guided task. Imaging the same region longitudinally will allow us to identify how cortical ensembles emerge or are remodeled as visual features become task-salient (Aim 1). Then, to test their functional role, I will artificially activate identified neural ensembles with two-photon holographic optogenetics to test whether their activity is sufficient to generate a percept and to elicit the expected visually-evoked response (Aim 2). Successful completion of these experiments will test whether neural ensembles are indeed building blocks of cortical function and will provide a new understanding of the link between perception and behavior. Furthermore, the current project will demonstrate how cortical ensembles integrate visual information with non-sensory variables during learning and will yield new inroads towards a more complete understanding of vision, a prerequisite to addressing the paucity of effective treatments options for illnesses such as blindness.